LLVM 12.0.0-rc2 has been tagged

Hi,

LLVM 12.0.0-rc2 has been tagged. Please test and report any issues in bugzilla.

-Tom

Uploaded Ubuntu 20.10 with lldb.

sha256sum clang+llvm-12.0.0-rc2-x86_64-linux-gnu-ubuntu-20.10.tar.xz
70d15e6284e43472c04d28c6fe9db39f6f46aceafd0e9f6fe2f73f4e2d235fdc

Failed Tests (7):
SanitizerCommon-asan-i386-Linux :: Linux/crypt_r.cpp
SanitizerCommon-asan-i386-Linux :: Posix/crypt.cpp
SanitizerCommon-lsan-i386-Linux :: Linux/crypt_r.cpp
SanitizerCommon-lsan-i386-Linux :: Posix/crypt.cpp
SanitizerCommon-ubsan-i386-Linux :: Linux/crypt_r.cpp
SanitizerCommon-ubsan-i386-Linux :: Posix/crypt.cpp
lldb-api :: tools/lldb-vscode/runInTerminal/TestVSCode_runInTerminal.py

Testing Time: 559.31s
Unsupported : 3416
Passed : 84218
Expectedly Failed: 267
Failed : 7

Previously reported bugs for the above failed tests
Bug 48946 - skipping incompatible /usr/lib/x86_64-linux-gnu/libcrypt.so when searching for -lcrypt
Bug 48960 - FAIL: lldb-api :: tools/lldb-vscode/runInTerminal/TestVSCode_runInTerminal.py

llvm-test-suite
Passed: 2405

Tests: 2825
Metric: exec_time

Program 12.0.0-rc1 12.0.0-rc2 diff
test-suite...ute/GCC-C-execute-pr56866.test 0.00 0.00 inf%
test-suite...te/GCC-C-execute-920501-6.test 0.00 0.00 inf%
test-suite...-C-execute-ieee-fp-cmp-8l.test 0.00 0.00 inf%
test-suite...-C-execute-ieee-fp-cmp-8e.test 0.00 0.00 inf%
test-suite...md_ops_test_op_pmaxud_285.test 0.00 0.01 inf%
test-suite...te/GCC-C-execute-990628-1.test 0.00 0.00 inf%
test-suite...te/GCC-C-execute-921013-1.test 0.00 0.00 inf%
test-suite...te/GCC-C-execute-990827-1.test 0.00 0.00 inf%
test-suite...C-execute-ieee-20000320-1.test 0.00 0.00 inf%
test-suite...te/GCC-C-execute-991016-1.test 0.00 0.00 inf%
test-suite...imd_ops_test_op_pminub_36.test 0.00 0.01 inf%
test-suite...te/GCC-C-execute-920728-1.test 0.00 0.00 inf%
test-suite...te/GCC-C-execute-991201-1.test 0.00 0.00 inf%
test-suite...te/GCC-C-execute-920501-5.test 0.00 0.00 inf%
test-suite...ute/GCC-C-execute-pr68328.test 0.00 0.00 inf%
Geomean difference nan%
12.0.0-rc1 12.0.0-rc2 diff
count 2807.000000 2807.000000 2779.000000
mean 172.137951 213.007611 inf
std 2732.775633 2922.118939 NaN
min 0.000000 0.000000 -1.000000
25% 0.000500 0.000400 -0.333333
50% 0.000600 0.000500 -0.047675
75% 0.012250 0.011650 0.333333

Neil Nelson

Windows is ready.

This time, the openmp runtime is included again (it didn't build for
RC1). I had to work around https://bugs.llvm.org/show_bug.cgi?id=49339
with a cmake flag, but besides that everything looks good.

$ sha256sum LLVM-12.0.0-rc2*.exe
ff1e23dfd98525439cc50b29f6f6fc6971ae703de25beab8085b6080ca090fc3
LLVM-12.0.0-rc2-win32.exe
5bca2ba3275277ad85dcd7a9c31346d93ac2d5909a8cc3c72289eef7fd326429
LLVM-12.0.0-rc2-win64.exe

build_llvm_1200-rc2.bat|attachment (5.08 KB)

Hi Tom,

Thanks for the notice. I'm a bit later than usual with running the Zig test suite against the LLVM release candidate this cycle, but I have now begun. Please give me a chance to file all the bugs that I find; last release cycle the Zig test suite identified 7 regressions in LLVM 11 release candidates.

Unfortunately, the first bug that I encountered is that LLVMDumpModule() is outputting syntax errors[1], making it impractical to report the other error that I have discovered. Furthermore, the other error I have discovered is blocking the full test suite from running, so I don't yet have an estimate of how many regressions have been identified.

Point being it may take some time and back-and-forths until all issues are identified and reported; I hope we can wait to cut the final tag until I get a chance to make sure the whole test suite passes.

[1]: https://bugs.llvm.org/show_bug.cgi?id=49358

Hi,

I’ve uploaded the windows on arm binaries:
cc0d601791405bab498b3acfc458f6ba40df4a86e91702f0ce1565bacacf0d77 LLVM-12.0.0-rc2-woa64.exe

This was built from the rc2 source snapshot without any patches on top \o/
Credit goes to my colleague Maxim Kuvyrkov for producing and testing the binaries.

Cheers,
Diana

Hi,

I’ve uploaded ARM and AArch64 binaries, md5sum and size are:
abad82d00cbc35f1b229483326842185 367M clang+llvm-12.0.0-rc2-aarch64-linux-gnu.tar.xz
7f328cd417452b5f15eab7cdfe38d15b 338M clang+llvm-12.0.0-rc2-armv7a-linux-gnueabihf.tar.xz

AArch64 contains openmp this time (was disabled in rc1) and has one new failure in compiler-rt:
libFuzzer :: fuzzer-finalstats.test (number of executed unit is 34 instead of 33)

ARM has the same LLD cross-dso failures reported in rc1.

Cheers,
Yvan

For 12.0.0 rc2, I have built and tested on both FreeBSD 11 and 12. No
additional patches were needed anymore (i.e. fixes for both issues I
found in rc1 were merged).

Main results on amd64-freebsd11:

  Unsupported : 5562 (rc1: 5557)
  Passed : 80165 (rc1: 80122)
  Expectedly Failed : 247 (rc1: 246)
  Timed Out : 6 (rc1: 2)
  Failed : 94 (rc1: 95)
  Unexpectedly Passed: 2 (rc1: 2)

Test suite results on amd64-freebsd11:

  Passed: 2399 (rc1: 2399)
  Failed: 3 (rc1: 3)

Main results on amd64-freebsd12:

  Unsupported : 5562 (rc1: 5557)
  Passed : 80172 (rc1: 80125)
  Expectedly Failed : 243 (rc1: 242)
  Timed Out : 4 (rc1: 4)
  Failed : 89 (rc1: 90)
  Unexpectedly Passed: 6 (rc1: 6)

Test suite results on amd64-freebsd12:

  Passed: 2399 (rc1: 2399)
  Failed: 3 (rc1: 3)

Main results on i386-freebsd11:

  Unsupported : 3976 (rc1: 3971)
  Passed : 76713 (rc1: 76669)
  Expectedly Failed : 224 (rc1: 223)
  Timed Out : 1 (rc1: 0)
  Failed : 52 (rc1: 51)
  Unexpectedly Passed: 1 (rc1: 1)

Main results on i386-freebsd12:

  Unsupported : 3976 (rc1: 3971)
  Passed : 76718 (rc1: 76672)
  Expectedly Failed : 222 (rc1: 221)
  Failed : 48 (rc1: 48)
  Unexpectedly Passed: 3 (rc1: 3)

Uploaded:
SHA256 (clang+llvm-12.0.0-rc2-amd64-unknown-freebsd11.tar.xz) = 4ab6e35a6c36931dd47615f807020657d037cc468ba1347e868430b334f01cd7
SHA256 (clang+llvm-12.0.0-rc2-amd64-unknown-freebsd12.tar.xz) = a413da1b4c1f0d34038b3ad6a91e80af4fe27aa1c7400b7f065f9844db017a6e
SHA256 (clang+llvm-12.0.0-rc2-i386-unknown-freebsd11.tar.xz) = cabf6006a0f2a573bc5fedb1a1f3ff0a803149e36e92847b2726daf2723f2ca6
SHA256 (clang+llvm-12.0.0-rc2-i386-unknown-freebsd12.tar.xz) = c3a701051d9d75dad96dbb31caaee40986573a91e4b033cb5db3ffa5cf06b2c4

-Dimitry

Hi,

PowerPC binaries uploaded,
Ubuntu SHA=69e97a92d5ea9ba1b9c29ba1de82069e29c4d22a
RHEL SHA=301da40b3265334555f4859b55b13fd140941724

Albion

clang+llvm-12.0.0-rc2-powerpc64le-linux-rhel-7.4.sha1 (98 Bytes)

clang+llvm-12.0.0-rc2-powerpc64le-linux-ubuntu-18.04.sha1 (102 Bytes)

Tom,

Sorry for being late on the ball here - but I have been traveling.
Testing on macOS seems to have yielded new failures from 11.0.1:

FAIL: AddressSanitizer-x86_64-darwin ::
TestCases/Posix/lto-constmerge-odr.cpp (6487 of 81886)
FAIL: AddressSanitizer-x86_64h-darwin ::
TestCases/Posix/lto-constmerge-odr.cpp (6979 of 81886)
FAIL: libomp ::
ompt/loadtool/tool_available_search/tool_available_search.c (79071 of
81886)
FAIL: libomp :: tasking/hidden_helper_task/gtid.cpp (79162 of 81886)
FAIL: libunwind :: signal_frame.pass.cpp (79248 of 81886)

The libunwind issue was there in the 11.x release but the address
sanitizer ones are new. I filed bugs for them here:

https://bugs.llvm.org/show_bug.cgi?id=49487 (asan)
https://bugs.llvm.org/show_bug.cgi?id=49488 (openmp)

not sure if you want to mark these as blockers.

The binaries are uploaded with the SHA-256:
d6da99de85169e3b02eff5dc9639f5791a476df1b1cdb3818ccd59002148135b